ABSTRACT Chemical exchange saturation transfer (CEST) imaging is an important molecular MRI technique that can generate contrast based on the proton exchange between labile protons in solutes and bulk water protons. Currently, CEST MRI holds great promise for numerous clinical molecular imaging applications. Notably, amide proton transfer (APT) imaging, a variant of the CEST-based molecular MRI technique, is based on the chemical exchange between free bulk water protons and the amide protons (-NH) of mobile proteins and peptides. Recent research in patients with brain tumors has shown that APT-weighted (APTw) MRI has the potential to enhance the noninvasive identification of brain tumors from peritumoral edema or normal tissue, to differentiate high- grade from low-grade tumors, or to differentiate treatment-related damage from tumor recurrence. Currently, however, APTw-MRI typically requires a long scan time due to the use of multiple radiofrequency (RF) saturation frequencies, possible multiple acquisitions to increase the signal-to-noise ratio (SNR), and a long RF saturation pulse (or pulse train), all of which limit clinical translation as the tumor patient is already undergoing many different MRI scans. The long-term goal of our proposed study is to develop a highly accelerated 3D APTw imaging technique to detect brain tumors and assess tumor response to therapy, which can easily be added to current clinical practice. The overall hypothesis is that the use of a blind compressed sensing (BCS) technique can significantly reduce APTw imaging scan time while maintaining a high degree of diagnostic image quality, which will further enhance the translation of this promising approach into a routine diagnostic tool. Our hypothesis will be tested through two specific aims: 1) to develop and optimize a highly accelerated 3D APTw-MRI method based on the BCS framework; 2) to validate this technology and assess the diagnostic accuracy of accelerated APTw images for the characterization of brain tumors in a clinical setting. In Aim 1, we will develop a 3D APTw imaging sequence with Cartesian pseudo-random k-space acquisition and a BCS reconstruction algorithm. We will also evaluate the efficacy of the BCS reconstruction in a quantitative manner, relative to conventional acceleration approaches, such as low-resolution acquisition and parallel imaging. In Aim 2, we will apply the technology developed in Aim 1 to patients with suspected treatment necrosis versus tumor recurrence, with the goal of validating the diagnostic accuracy of this accelerated scan to distinguish between radiation necrosis and tumor recurrence on a 3 T clinical scanner. The innovations of the proposed study are in the development of a 3D APTw-MRI pulse sequence for pseudo-random k-space undersampling acquisition, BCS image reconstruction from the highly undersampled data, and the use of the novel APTw methods in guiding stereotactic biopsy. This work will help to facilitate the clinical translation of APTw-MRI into a routine diagnostic tool to monitor tumor response and progression of therapy, and has applications to other diseases, such as hyperacute stroke and other neurological disorders.